The human CNOT1-CNOT10-CNOT11 complex forms a structural platform for protein-protein interactions

Abstract

The evolutionary conserved CCR4-NOT complex functions in the cytoplasm as the main mRNA deadenylase in both constitutive mRNA turnover and regulated mRNA decay pathways. The versatility of this complex is underpinned by its modular multi-subunit organization, with distinct structural modules actuating different functions. The structure and function of all modules are known, except for that of the N-terminal module. Using different structural approaches, we obtained high-resolution data revealing the architecture of the human N-terminal module composed of CNOT1, CNOT10, and CNOT11. The structure shows how two helical domains of CNOT1 sandwich CNOT10 and CNOT11, leaving the most conserved domain of CNOT11 protruding into solvent as an antenna. We discovered that GGNBP2, a protein identified as a tumor suppressor and spermatogenic factor, is a conserved interacting partner of the CNOT11 antenna domain. Structural and biochemical analyses thus pinpoint the N-terminal CNOT1-CNOT10-CNOT11 module as a conserved protein-protein interaction platform.

Keywords: AlphaFold; CCR4-NOT complex; CP: Molecular biology; GGNBP2; X-ray crystallography; cryo-EM; deadenylation; mRNA decay; protein-protein interactions.

Graphical abstract

Figure 1

Structure of the N-terminal module of the human CCR4-NOT complex: A structured core with an antenna domain (A) Coomassie-stained SDS-PAGE analysis of the purified recombinant sample of the human CNOT1_N-CNOT10-CNOT11 complex used for crystal structure determination. The complex corresponds to the N-terminal module of CCR4-NOT and includes only the N-terminal region of the scaffold protein CNOT1_N. (B) Schematic representation of the domain organization of human CNOT1_N (yellow), CNOT10 (light green), and CNOT11 (blue). The structured domains are indicated as rectangles and linker regions are shown as black lines. Labeling of the different portions of the proteins corresponds to the structural analysis, as discussed in the text. In particular, the N-terminal, middle, and C-terminal region of CNOT11 are indicated with the corresponding suffix N, M, C. HEAT, TPR, MIF4G; C9BD (CNOT9 binding domain). (C) Crystal structure of the CNOT1_N-CNOT10-CNOT11 complex in two different orientations, colored and labeled as in (B). The model of the structured core (CNOT1_N-CNOT10-CNOT11_M-N) is refined to 3.1 Å resolution. Due to the apparent flexible connection to the core, the model of the “antenna” domain (CNOT11_C) was determined separately at 2.2 Å resolution and fitted in the density of the ternary complex.

Figure 2

The structured core of the CNOT1_N-CNOT10-CNOT11 complex is formed by evolutionary conserved interactions (A–F) Zoom-in views of the major structural interactions in the CNOT1_N-CNOT10-CNOT11_N-M core of the CCR4-NOT N-terminal module. The individual views are also indicated in the context of the entire structure. (A–C) highlight conserved interactions within the inner layer of the complex comprising the extended CNOT11_M region and the CNOT10 TPR superhelix. The interactions with the two outer layers are shown in (D and E) (CNOT1_N domain 1, CNOT11, and CNOT10) and (F) (CNOT1_N domain 2 and CNOT10). The evolutionary conservation of the interactions is shown in Figures S2A–S2C. (G) Biochemical validation of the structural analysis. Co-immunoprecipitation of endogenous CNOT10 with GFP-tagged CNOT11 truncated proteins. HEK293 cells were transfected with plasmids expressing GFP-tagged human CNOT11 fusion proteins or empty GFP expression vector. Proteins were immunoprecipitated with GFP-Trap magnetic beads and the co-precipitation was analyzed by western blotting. For some constructs, limited CNOT11 degradation during immunoprecipitation generated additional lower molecular-weight bands.

Figure 3

The antenna domain of CNOT11 mediates the recruitment of GGNBP2 to CCR4-NOT (A) CNOT11_C protein partners from a human placenta cDNA library identified in a yeast two-hybrid screen. The number of clones encoding fusions with different ORF is shown. The common region found in the different clones encoding CNOT11_C partners is: BRAP (113–490), EDRF1 (283–595), FAM193A (1,155–1,240), FAM193B (680–822), GGNBP2 (483–697), and GPBP1L1 (307–462). An arrow marks GGNBP2 also identified by affinity purification-mass spectrometry. (B) Identification of proteins co-purifying with the N-terminus of CNOT11 by mass spectrometry. HEK293 cells were transfected with plasmids expressing TAP-tagged CNOT11 fusion proteins that were purified by two tandem affinity purification steps. Co-purified proteins were identified by mass spectrometry. Proteins found in control purification (such as keratins) were filtered out. The average number of Peptide-Spectrum Match (PSM) and standard deviation for the top 14 identified proteins in three mass spectrometry replicates for CNOT11_N-M-C and CNOT11_N-M are shown. (C) Identification of proteins co-purifying with the C-terminus of CNOT11 by mass spectrometry. Protocol similar to (B). The average number of Peptide-Spectrum Match (PSM) and standard deviation for the top 6 identified proteins in three mass spectrometry replicates for CNOT11_N-M-C and CNOT11_C are shown. An arrow marks GGNBP2, also identified in the two-hybrid screen. (D) Co-immunoprecipitation of endogenous CNOT7, CNOT10, and CNOT11 with GFP-tagged GGNBP2 proteins. HEK293 cells were transfected with plasmids expressing GFP-tagged human GGNBP2 fusion proteins (numbers in parentheses indicate amino acids included, FL = Full-Length), or empty GFP expression vector. Proteins were immunoprecipitated with GFP-Trap magnetic agarose beads and the co-precipitation was analyzed by western blotting on 8% PAGE. Note that CNOT11 always appears as multiple bands in cell lysates while overexpressed GFP-GGNBP2 often appears slightly degraded. (E) Identification of proteins co-purifying with GFP-GGNBP2 by mass spectrometry. The plot compares the average number of PSM for each protein obtained in purification with GFP-GGNBP2FL (y axis) versus control purifications with GFP alone and GFP-GGNBP2(1-482) (x axis). GGNBP2 is indicated in red while subunits of the CCR4-NOT complex and partners are indicated in blue.

Figure 4

The CNOT11 antenna domain recognizes the C-terminal segment of GGNBP2 (A) Structure-based sequence alignment of CNOT11-binding domain of GGNBP2. Highlighted in dark and light orange are residues with high and medium conservation across the species shown. A schematic representation shows the position of the two α-helices with respect to the amino acid sequence. (B) Co-immunoprecipitation of endogenous CNOT11 with GFP-tagged truncated GGNBP2 proteins. HEK293 cells were transfected with plasmids expressing GFP-tagged human GGNBP2 fusion proteins (numbers in parentheses indicate amino acids included, FL = Full-Length, GGNBP2 C-terminal residues were replaced by an HA tag in the GFP-GGNBP2(1-625)-HA construction). Proteins were immunoprecipitated with GFP-Trap magnetic agarose beads and the co-precipitation was analyzed by western blotting. Note that CNOT11 always appears as multiple bands in cell lysates. (C) Reconstitution of the CNOT11-GGNBP2 interaction in vitro with recombinant proteins. Chromatogram from size-exclusion chromatography and corresponding Coomassie-stained SDS-PAGE analysis of the peak fraction shows the presence of a complex between CNOT11_C and two C-terminal segments of GGNBP2: residues 626–697 and the truncation mutant encompassing residues 638–673 that was still able to interact with CNOT11_C and that was used for the crystal structure determination of the complex. (D) Overall structure of the complex between CNOT11_C (in blue) and GGNBP2_C (in orange, the atomic model from the crystal structure; in red, superposed the model from AlphaFold predictions, Figure S4A). The HEAT repeats of CNOT11_C are labeled, as well as the GGNBP2_C helices (α0 and α1). On the right, the CNOT11_{C -} GGNBP2_C complex (atomic model from the crystal structure) is positioned in the CNOT1_N-CNOT10-CNOT11 module.

Figure 5

The GGNBP2 C-terminal segment wraps around the CNOT11 antenna domain with conserved interactions (A) Zoom-in views of major structural interactions in the CNOT11_C-GGNBP2_C complex. Labeled by asterisks in the top right panel are the CNOT11 residues Ala476 and Phe491 and the GGNBP2 residues Ile643, Phe651, and Phe658 that are mutated in the assays shown in (B–D). (B) Interaction of CNOT11 wild-type or mutant (A476E, F491D) proteins with GGNBP2 in yeast two-hybrid assays. The host yeast strain was co-transformed with plasmids expressing the indicated LexA-DNA-Binding-Domain (DBD) and Gal4-Activating-Domain (AD) fusion proteins (numbers in parentheses indicate included amino acids; FL, full-length; WT, wild-type). Interaction between the different chimeric proteins indicated was assessed by β-galactosidase assays performed in duplicate. Activities are expressed in arbitrary units and plotted on a logarithmic scale. Three biological replicates were assayed and the average activity was plotted with standard deviation indicated as error bars. (C) Interaction of GGNBP2 wild-type or mutant (I643A and/or F651R, F658D) proteins with CNOT11 in yeast two-hybrid assays. Protocol as in (B). (D) Co-immunoprecipitation of wild-type (WT) and mutant CNOT11 with WT and mutant GGNBP2 proteins. HEK293 cells were co-transfected with plasmids expressing GFP-tagged human WT and mutant GGNBP2 proteins with TAP-tagged CNOT11 WT and mutant proteins (numbers in parentheses indicate amino acids included). Proteins were immunoprecipitated with GFP-Trap magnetic M-270 beads and the co-precipitation was analyzed by western blotting on 10% PAGE.